Statistics pitfalls

Statistical Science Avoiding Statistical Pitfalls (p. 242) Volume 6, Number 3, August 1991... 

The BLOGP model is questionable for its large number of parameters and especially for the use of four squared terms along with the three terms at the fourth power. The only justification for such high power terms is statistical, because no intermolecular interactions can offer a physicochemical rationale. Moreover, a careful analysis of Eq. [27] reveals two severe statistical pitfalls ... 

Boolean descriptors. These descriptors are easy to derive but their introduction in dummy regression analysis can yield statistical pitfalls when they are used for predictive purposes. To overcome this problem, it is possible to use the stochastic regression analysis. This method consists of performing, as a first step, a correspondence factor analysis (CFA) and then carrying out a regression analysis from the CFA factors. The stochastic regression analysis can also be used when the molecular descriptors are frequencies of occurrence. ... 

Examples ofTypical QRA Objectives Types of Eacility Resotirces/Personnel Classical Limitations of QRA Example of Mortality Statistics Issues Affecting Perception of Risk Typical Pitfalls in Using QRA... 

Although anti-TNFa agents are well tolerated and have a good overall safety profile, pitfalls to the use of these drugs apparent with increasing clinical experience include infective complications and, in particular, reactivation of tuberculosis. To date, no statistically significant increased rate of tumour occurrence over that expected has been noted although cases of lymphoma have rarely been reported in patients treated with TNFa blockade. 

The large deformability as shown in Figure 21.2, one of the main features of rubber, can be discussed in the category of continuum mechanics, which itself is complete theoretical framework. However, in the textbooks on rubber, we have to explain this feature with molecular theory. This would be the statistical mechanics of network structure where we encounter another serious pitfall and this is what we are concerned with in this chapter the assumption of affine deformation. The assumption is the core idea that appeared both in Gaussian network that treats infinitesimal deformation and in Mooney-Rivlin equation that treats large deformation. The microscopic deformation of a single polymer chain must be proportional to the macroscopic rubber deformation. However, the assumption is merely hypothesis and there is no experimental support. In summary, the theory of rubbery materials is built like a two-storied house of cards, without any experimental evidence on a single polymer chain entropic elasticity and affine deformation. 

Linearization is here defined as one or more transformations applied to the X- and/or y-coordinates in order to obtain a linear y vs. x relationship for easier statistical treatment. One of the more common transformations is the logarithmic one it will nicely serve to illustrate some pitfalls. 

A probabilistic risk assessment (PRA) deals with many types of uncertainties. In addition to the uncertainties associated with the model itself and model input, there is also the meta-uncertainty about whether the entire PRA process has been performed properly. Employment of sophisticated mathematical and statistical methods may easily convey the false impression of accuracy, especially when numerical results are presented with a high number of significant figures. But those who produce PR As, and those who evaluate them, should exert caution there are many possible pitfalls, traps, and potential swindles that can arise. Because of the potential for generating seemingly correct results that are far from the intended model of reality, it is imperative that the PRA practitioner carefully evaluates not only model input data but also the assumptions used in the PRA, the model itself, and the calculations inherent within the model. This chapter presents information on performing PRA in a manner that will minimize the introduction of errors associated with the PRA process. 

Multidimensional linear regression analysis is the most often employed statistical method for QSARs, This popularity is coupled to the acceptance of the Hansch method for QSAR analyses. The techniques and pitfalls of regression analysis have been well described.(44,45)... 

Analysis methods for electrochemical noise data can be separated into three categories, (1) deterministic, (2) statistical, and (3) spectral. Deterministic methods involve the use of mixed potential theory to explain the oscillations that occur. For example, if the ZRA current increases suddenly while the potential difference between the two current electrodes and the potential electrode increases, localized corrosion has likely initiated on one of the current electrodes. A common pitfall in such a measurement is that if a nominally identical reference electrode is used, it could pit as well, leading to no change in potential versus the coupled electrodes. Due to the need for careful interpretation, deterministic methods are not widely used. 

Rietschell RL (1982) Advances and pitfalls in irritant and allergen testing. J Soc Cosmet Chem 33 309-313 Wooding WH, Opdyke DL (1967) A statistical approach to the evaluation of cutaneous responses to irritants. J Soc Cos-met Chem 16 809-829... 

Linear regression analysis has pitfalls. There is always the possibility of chance correlations. Hence, we opted to analyze the data using an alternate statistical method, namely cluster analysis. The data were scaled so that each of the descriptors ranged in value between 0 and 1. Minimal tree spanning methods was employed in the determination of clusters (24). 

Also non-linear regression, that is, using quadratic terms such as (logP) and cross terms, may be used. However, as described in detail, there are a number of pitfalls to this method. A statistically more robust method which could be used instead of MLR is the PLS regression method. 

Some statistics concepts such as mean, range, and variance, test of hypothesis, and Type I and Type II errors are introduced in Section 2.1. Various univariate SPM techniques are presented in Section 2.2. The critical assumptions in these techniques include independence and identical distribution [iid) of data. The independence assumption is violated if data are autocorrelated. Section 2.3 illustrates the pitfalls of using such SPM techniques with strongly autocorrelated data and outlines SPM techniques for autocorrelated data. Section 2.4 presents the shortcomings of using univariate SPM techniques for multivariate data. 

D. Murray, J. Schwartz, S. Lichter. It Ain t Necessarily So. New York Rowman and Littlefield, 2001. (Note This is an excellent reference which illustrates many pitfalls of statistics when applied to demographics and other issues.) P. Schulze, J. Mealy. Population Growth, Technology and Tricky Graphs. American Scientist 89 209-211. 

This chapter describes some of the approaches and techniques used currently to derive in silico models for the prediction of absorption, distribution, metabolism, elimination/excretion, and toxicity (ADMET) properties. The chapter also discusses some of the fundamental requirements for deriving statistically sound and predictive ADMET relationships as well as some of the pitfalls and problems encountered during these investigations. It is the intention of the authors to make the reader aware of some of the challenges involved in deriving useful in silico ADMET models for drug development. 

This chapter has illustrated several identification methods that are used to determine dynamic parameters or models from experimental plant data. The simple and effective relay feedback test is a powerful tool for practical identification if the objective is the design of feedback controllers. The more complex and elegant statistical methods are currently popular with the theoreticians, but they require a very large amount of data (long test periods) and their effective use requires a high level of technical e q3ertise. It is very easy to get completely inaccurate results from these sophisticated tests if the user is not aware of all the potential pitfalls (both fundamental and numerical). 

---